INDEX

Explanations

to prove or demonstrate

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 candid

-0.12

avit

-0.12

 spont

-0.12

-0.11

alem

-0.11

 reluctant

-0.11

ty

-0.11

ivet

-0.11

irit

-0.11

rit

-0.10

POSITIVE LOGITS

 prove

0.41

 proving

0.40

 proves

0.34

prove

0.31

 proved

0.29

 demonstrating

0.26

 impress

0.25

è¯ģ

0.25

 demonstrate

0.24

 proofs

0.21

Activations Density 0.185%